Pronunciation
Mandarin has 4 main tones, plus a neutral (unstressed) tone. The same syllable, pronounced with a different tone, is a completely different word — tones aren't an accent or a nuance, they're part of the word's identity, exactly like a different vowel would be in English.
| Tone | Mark | Contour | Example |
|---|---|---|---|
| 1st | ā | Flat and high, no variation ("held on a single note") | 妈 mā — mother |
| 2nd | á | Rising, from mid to high (like a question in English) | 麻 má — hemp |
| 3rd | ǎ | Dips down then rises slightly (usually reduced to just the dip in natural speech, see below) | 马 mǎ — horse |
| 4th | à | Sharp, fast fall from high to low (like a curt command) | 骂 mà — to scold |
| Neutral | a (no mark) | Short, unstressed, pitch depends on the syllable before it | 吗 ma — question particle |
Four completely unrelated words, same syllable "ma" — the tone alone tells them apart. This is the single example every Mandarin course starts with, for good reason: it makes the stakes of getting tones right immediately obvious.
In pinyin, the tone is marked with a small sign (the mark from the table above) placed on a single vowel of the syllable, even when there's more than one. The usual priority rule: a always takes the mark if present (hǎo, not haǒ); otherwise it's o or e; otherwise, in the combinations iu and ui, the second vowel carries it (liù, duì). The neutral tone never carries a mark — its absence is itself meaningful: it signals a short, unstressed syllable, not a "default" tone identical to the 1st tone.
The tone marked in pinyin isn't always the one actually pronounced once a word sits inside a sentence — Mandarin has several automatic, predictable tone-change rules, not isolated exceptions to memorize case by case.
Two 3rd tones in a row
The first of the two is actually pronounced as a 2nd tone — a native speaker never fully dips twice in a row. 你好 (nǐ hǎo) is therefore really pronounced "ní hǎo", even though the pinyin spelling keeps nǐ. Before a 1st, 2nd, 4th, or neutral tone, a 3rd tone instead loses its final rise and reduces to a simple low dip (the "half 3rd tone").
不 (bù, "not") before a 4th tone
不 shifts from 4th tone (bù) to 2nd tone (bú) right before another 4th tone: 不是 is pronounced búshì, 不对 búduì, 不要 búyào. Before any other tone, 不 keeps its original form: 不好 bù hǎo, 不来 bù lái, 不走 bù zǒu.
一 (yī, "one") depending on the word that follows
一 stays 1st tone on its own, at the end of a word, or as an ordinal number (第一 dì-yī, "first"; 星期一 xīngqīyī, "Monday"). Before a 4th tone, it becomes yí: 一个 yí ge, 一样 yíyàng, 一定 yídìng. Before a 1st, 2nd, or 3rd tone, it becomes yì: 一天 yìtiān, 一年 yìnián, 一起 yìqǐ.
Good news for practice: these three rules are automatic once you know the tone of the following word — recognizing the pattern by ear matters far more early on than trying to calculate it yourself in real time mid-conversation.
| Syllable | Word 1 | Word 2 | Word 3 | Word 4 |
|---|---|---|---|---|
| tang | 汤 tāng — soup | 糖 táng — sugar | 躺 tǎng — to lie down | 烫 tàng — scalding hot |
| mai | — | — | 买 mǎi — to buy | 卖 mài — to sell |
| wen | — | — | 吻 wěn — to kiss | 问 wèn — to ask |
| shi | — | 十 shí — ten | — | 是 shì — to be |
| xiang | — | — | 想 xiǎng — to want, to think | 向 xiàng — towards |
汤/糖/躺/烫 is a great one to say out loud: all 4 tones on the exact same syllable, from hot liquid (汤) to being scalding hot (烫), passing through sugar (糖) and lying down (躺) — no relation in meaning between them, tone alone tells them apart.
Flattening tones once a sentence speeds up
The 3rd tone especially often ends up pronounced flat instead of actually dipping — particularly outside of sandhi (see above), where its full shape needs to stay audible.
Mixing up the shape of the 2nd and 3rd tone
Both sound like they "rise" to a beginner's ear, but the 2nd rises the whole way through while the 3rd dips first before rising (or just stays low, under sandhi) — a difference in shape, not just pitch height.
Treating the neutral tone as a "default tone"
A neutral-tone syllable (like 吗 or 的) has no fixed pitch of its own: it borrows its color from the preceding tone, and is always short and unstressed — over-stressing it makes it sound like a 1st tone, which changes what the listener understands.
Reading about tones only goes so far — the skill is auditory and comes from repeated exposure with immediate correction. HSK Trainer has a dedicated tone drill: the character, the toneless pinyin and the meaning are shown, and you pick the correctly-accented pinyin among 4 options (3 of them with the tones deliberately shuffled) — built to isolate exactly the skill described above.
How many tones does Mandarin have?
Four main tones, plus a neutral tone sometimes counted as a fifth — 5 distinct pitches in speech, even though only 4 are traditionally numbered.
Does getting a tone wrong actually matter?
Yes: two words with the same syllable but different tones are two completely unrelated words. A wrong tone can make a sentence incomprehensible, not just "sound foreign".
What is tone sandhi?
The automatic, predictable change of a word's tone based on the tone that follows it — see the dedicated section above for the 3 most common rules (3rd+3rd, 不, 一).
Can tones be learned just by listening?
Partly, but slowly without immediate corrective feedback: the ear improves mostly when a mistake is actively flagged, not from passive listening alone.